GmPAO-mediated polyamine catabolism enhances soybean Phytophthora resistance without growth penalty

Abstract

Plant immunity is activated upon perception of pathogens and often affects growth when it is constitutively active. It is still a challenge to balance plant immunity and growth in disease resistance breeding. Here, we demonstrated that soybean (Glycine max) polyamine oxidase (GmPAO) confers resistance to multiple Phytophthora pathogens, but has no obvious adverse impact on agronomic traits. GmPAO produces H₂O₂ by oxidizing spermidine and spermine. Phytophthora sojae induces an increase in these two substrates, and thus promotes GmPAO-mediated polyamine catabolism specifically during infection. Interestingly, we found that the two substrates showed higher accumulation in transgenic soybean lines overexpressing GmPAO than in WT and CK after inoculation with P. sojae to ensure H₂O₂ production during infection, rather than directly inhibit P. sojae. In these transgenic soybean plants, the significantly enhanced resistance to different P. sojae isolates was achieved; PAMP-induced H₂O₂ accumulation was enhanced by GmPAO overexpression. Moreover, transient expression of GmPAO also significantly improved Nicotiana benthamiana resistance to Phytophthora capsici and Phytophthora parasitica in agroinfiltration assays. Our results provide a novel approach to allow rapid defense responses in plants upon pathogen infection while minimizing growth penalties under normal conditions, with a clear mechanism in which plant promotes H₂O₂ production via pathogen-activated substrates.

Background

The oomycete genus Phytophthora, which consists of nearly 200 described species, has been and continues to be a significant threat to global agroecosystems and food security (Kamoun et al. 2015; Wang et al. 2020). For example, Phytophthora sojae in soybean causes ‘damping off’ at the seedling stage and root rot in adult plants, resulting in an annual loss of $1–2 billion worldwide (Zhou et al. 2021). Another prevalent phytopathogenic oomycete, P. capsica, is a highly dynamic and destructive pathogen attacking vegetables, including cucurbit, pepper, tomato, eggplant, lima bean, etc. (Lamour et al. 2012; Li et al. 2019b). Furthermore, P. parasitica causes notable diseases of brown rot, foot rot, and black shank in tobacco, affecting almost all tobacco growing areas worldwide with losses as high as 100% (Panabieres et al. 2016; Yu et al. 2020). Accordingly, plants respond to pathogen infection using an efficient, dynamic, and hierarchical immune system (Jones and Dangl 2006; Wang et al. 2020). Immune activation in plants confers resistance against invading pathogens, but constitutive immune responses are costly and impede plant growth and environmental fitness (Deng et al. 2020). Therefore, a delicate balance between warding off potential pathogens and avoiding negative impacts on growth fitness must be maintained (Ning et al. 2017; Gao et al. 2021). Genetic engineering is a desirable approach to solve these problems via expression of carefully selected genes (Deng et al. 2020; Liu et al. 2021).

So far, some emerging breeding approaches for crop disease resistance have been extensively studied, including (1) molecular strategies for enhancing plant immune system (Frailie and Innes 2021); (2) modification of plant susceptibility genes (Tian et al. 2020; Koseoglou et al. 2022); (3) engineering antimicrobial proteins (Sinha and Shukla 2019; Liu et al. 2021); (4) interfering with pathogenicity (Li et al. 2021); and (5) manipulation of the expression of ‘master-switch’ genes, such as kinases, transcription factors, and signal molecules (Ma et al. 2017; Wang et al. 2018; Narvaez et al. 2020; Liu et al. 2021). At the same time, several strategies have been developed to manipulate the expression of ‘master-switch’ genes. For example, uORFs-mediated translational control of Arabidopsis NPR1 generated broad-spectrum resistance without the NPR1-induced growth penalty in transgenic rice (Xu et al. 2017). Balancing yield and immunity has also been greatly achieved. For example, a rice locally adaptive ROD1 allele confers higher basal resistance without loss of crop yield (Gao et al. 2021), whereas IPA1 promotes both yield and disease resistance by sustaining a balance between growth and immunity (Wang et al. 2018).

Dissection of plant–pathogen interactions at molecular level uncovers potential engineering targets for plant defense enhancement. As is well known, hydrogen peroxide (H₂O₂) is a critical signal triggering plant defense response and can be rapidly produced under biotic stresses (Qi et al. 2018). Plant polyamines (PAs) are involved in immunity against pathogens, notably by amplifying pattern-triggered immunity (PTI) responses through the production of H₂O₂ (Gerlin et al. 2021). H₂O₂ can be generated via spermidine (Spd) and spermine (Spm) oxidation catalyzed by PA oxidase (PAO) (Gerlin et al. 2021). Although previous studies found that overexpression of PAO can improve disease resistance in model plants Arabidopsis (Mo et al. 2015) and tobacco (Moschou et al. 2009), no transgenic application of PAO-encoding genes in crops has been reported.

Here, we aim to improve Phytophthora resistance in soybean by overexpressing its native PAO (GmPAO) gene that oxidizes Spd and Spm to generate H₂O₂. We demonstrated that the two substrates, Spd and Spm, were significantly increased upon P. sojae infection. Transgenic soybean lines overexpressing GmPAO gene exhibited higher H₂O₂ production only in response to P. sojae infection. The elevated PTI activation was also observed in the GmPAO overexpression soybean lines. Moreover, Spd and Spm contents were highly induced in these transgenic lines. As expected, the transgenic soybean lines confer disease resistance to P. sojae without growth penalty. Thus, we propose a method by expressing plant PAO genes to induce disease resistance while minimizing cellular damage caused by H₂O₂.

Results

Spd and Spm contents are increased in soybean after infection with P. sojae

The major polyamines (PAs) in plants are the diamine putrescine (Put), the triamine spermidine (Spd), and the tetraamine spermine (Spm) (Seifi and Shelp 2019). As PAs have an important role in the interaction between plants and pathogens (Gerlin et al. 2021), we determined to investigate PA changes in soybean plants in response to P. sojae infection using ultra-high-performance liquid chromatography coupled to quadrupole Orbitrap high-resolution mass spectrometry (UHPLC-MS) method. The results showed that the Put content in soybean was significantly decreased during P. sojae infection (Fig. 1a). In contrast, the Spd content was significantly increased to 3.0-fold at 16 h post-inoculation (hpi), 2.0-fold at 32 hpi, and 1.6-fold at 48 hpi (Fig. 1a). Furthermore, the Spm content in soybean was increased to 1.4-fold at 16 hpi, 2.0-fold at 32 hpi, and 1.3-fold at 48 hpi (Fig. 1a). Collectively, the results suggest that the Spd and Spm contents in soybean were increased during P. sojae infection.

Fig. 1

Transgenic soybean lines overexpressing GmPAO display significantly boosting H₂O₂ production at the P. sojae infection sites. WT: wild-type soybean cultivar Williams 82; CK: non-GmPAO transgenic soybean line. T1-1 (including T1-1-1, T1-1-3, T1-1-4, and T1-1-5) and T1-4-1 are progenies of two independent transgenic soybean lines overexpressing GmPAO gene. a Changes in the contents of putrescine (Put), spermidine (Spd), and spermine (Spm) in WT plants in response to P. sojae infection. Error bar represents mean ± SD, n = 3, n represents sample number. b, d PAO enzymatic activity and H₂O₂ content in transgenic soybean plants during P. sojae infection. RU: relative activity units (Error bar represents mean ± SD, n = 3, n represents sample number). c DAB staining of H₂O₂ at 16 hpi. Scale bar: 1 cm (upper). Micrograph of DAB staining at P. sojae infection sites. Scale bar: 20 μm (lower). All the experiments were performed three times (biological replicates) with similar results. The data in a, b, and d were analyzed by Shapiro–Wilk test to determine the Normality and Lognormality Tests across groups, and then analyzed by one-way ANOVA with post-hoc Dunnett's multiple comparisons test for groups that had passed the normality test (ns, no significant difference; *P < 0.05; **P < 0.01). The exact n, SD values, and P values are shown in the data

Transgenic soybean lines overexpressing GmPAO gene display significantly boosting hydrogen peroxide production at infection sites

We cloned the full-length cDNA sequence of GmPAO (Glyma.09G227500.1) according to the genome sequence of Glycine max Wm82.a2.v1 in the Phytozome database, and generated stable transgenic soybean lines overexpressing GmPAO gene. Two independent transgenic soybean lines (T1-1 and T1-4-1) overexpressing GmPAO were generated and characterized by molecular screening at the T5 generation (Additional file 1: Figure S1 and Additional file 2: Table S1). T1-1-1, T1-1-3, T1-1-4, and T1-1-5 sublines were all derived from T1-1. To further confirm the expression of GmPAO in the transgenic lines, we detected the PAO enzymatic activity. The results showed that the PAO enzymatic activity in transgenic soybean lines was increased to ~ 1.8-fold compared with that in WT and the non-GmPAO transgenic line (CK) in the absence of P. sojae challenge (Fig. 1b). Furthermore, we also detected the PAO enzymatic activity during P. sojae infection. The results showed that the PAO enzymatic activity in transgenic soybean lines was increased to 2.5-fold at 16 hpi, and 1.8-fold at both 32 and 48 hpi compared with that in WT and CK (Fig. 1b). Thus, in comparison with WT and CK, all the GmPAO overexpression soybean lines showed significantly enhanced polyamine synthetic capacity before and after P. sojae infection, with the highest PAO enzymatic activity being detected at 16 hpi.

To monitor the effect of GmPAO-mediated polyamine catabolism, we conducted a series of experiments to test H₂O₂ production. First, we investigated H₂O₂ accumulation in the transgenic soybean lines using 3,3′-diaminobenzidine (DAB) staining assay (Wang et al. 2020; Zhou et al. 2021). Compared with WT and CK, the five transgenic soybean lines showed significantly enhanced H₂O₂ accumulation at 16 hpi (Fig. 1c). Interestingly, we found that H₂O₂ accumulation occurred only at the P. sojae infection sites, on the contrary, there was no H₂O₂ accumulation at the P. sojae no-infection sites (Fig. 1c). These results preliminarily indicated that GmPAO-mediated H₂O₂ generation is activated upon P. sojae infection. H₂O₂ production is not only required in many important defense reactions (Qi et al. 2018) but also inevitable results of oxidative damage (Mittler 2017; Kerchev and Van Breusegem 2022). Intriguingly, GmPAO-mediated H₂O₂ accumulation only occurred at the P. sojae infection sites to minimize cellular damage caused by H₂O₂.

Second, we examined the relative H₂O₂ levels in WT, CK and transgenic soybean plants using H₂O₂ assay kit (Yin et al. 2018; Zhang et al. 2019a). The results showed that the H₂O₂ content in transgenic soybean lines was only weakly increased to 1.2-fold compared with that in WT and CK in the absence of P. sojae challenge (Fig. 1d), indicating a relatively normal H₂O₂ level in the transgenic soybean lines under normal conditions. We further detected the relative H₂O₂ level during P. sojae infection. The results showed that the relative H₂O₂ level in the transgenic soybean lines was increased to 2.3-fold at 16 hpi, 1.7-fold at 32 hpi, and 1.4-fold at 48 hpi compared with that in WT and CK (Fig. 1d). These results further indicated that GmPAO-mediated H₂O₂ is activated upon P. sojae infection.

Together, the results indicated that Spd/Spm contents in soybean plants were upregulated in response to P. sojae infection, and GmPAO-mediated Spd/Spm catabolism promoted H₂O₂ production mostly at the P. sojae infection sites.

GmPAO promotes PAMP-induced H₂O₂ accumulation

PAs are involved in PTI responses (Gerlin et al. 2021). We therefore investigated the effect of GmPAO on the pathogen-associated molecular patterns (PAMP)-induced reactive oxygen species (ROS) burst using an L-012 based assay (Pi et al. 2022). We examined the well-known bacterial pattern flg22-induced ROS and found that there was a significantly enhanced ROS production in the transgenic soybean lines (Fig. 2a). Chitin-induced ROS accumulation was also enhanced by GmPAO (Fig. 2b). Next, we validated the regulation activity of flg22/chitin-induced ROS accumulation by GmPAO using a transient expression assay in Nicotiana benthamiana. Consistent with the results in transgenic soybean lines, the transient expression of GmPAO in N. benthamiana enhanced flg22/chitin-induced ROS accumulation (Fig. 2c, d). These results suggest that GmPAO can promote flg22/chitin-induced H₂O₂ accumulation.

Fig. 2

GmPAO promotes PAMP (flg22 or chitin)-induced H₂O₂ accumulation. a, b GmPAO promoted PAMP (flg22 or chitin)-induced H₂O₂ accumulation in transgenic soybean lines. The leaves (2-week-old) were subjected to flg22- (a) or chitin-induced (b) ROS examination, and the relative luminescence unit (RLU) value was recorded. Error bar represents mean ± SD, n ≥ 8, n represents sample number. c, d GmPAO promoted PAMP (flg22 or chitin)-induced H₂O₂ accumulation in N. benthamiana. The indicated constructs were transiently expressed by Agrobacterium-mediated transient expression for 2 days and subjected to flg22- (c) or chitin-induced (d) ROS examination. GFP was used as a control. The protein expression is shown in the right panel. Error bar represents mean ± SD, n = 8, n represents sample number. All the experiments were performed three times (biological replicates) with similar results. The data were analyzed by Shapiro–Wilk test to determine the Normality and Lognormality Tests across groups, and then analyzed by unpaired t test for groups that had passed the normality test (**P < 0.01). The exact n, SD values, and P values are shown in the data

Transgenic soybean lines overexpressing GmPAO gene display significantly elevated contents of Spd and Spm when challenged by P. sojae

To investigate the effect of GmPAO-mediated polyamine catabolism, we measured the contents of PAs in transgenic soybean lines upon infection by P. sojae. The results showed that there was no significant change in Spd content among the transgenic soybean lines, WT, and CK in the absence of P. sojae challenge (Fig. 3a). Furthermore, we detected the Spd content during P. sojae inoculation. The results showed that the Spd content in transgenic soybean lines was increased to 1.7-fold at 16 hpi, 1.5-fold at 32 hpi, and 1.3-fold at 48 hpi compared with that in WT and CK (Fig. 3a). Thus, in comparison with WT and CK, all GmPAO overexpression soybean lines showed significantly elevated Spd content after P. sojae infection.

Fig. 3

Changes in the contents of Spd (a) and Spm (b) in transgenic soybean lines upon P. sojae inoculation. Error bar represents mean ± SD, n = 3, n represents sample number. All the experiments were performed three times (biological replicates) with similar results. The data were analyzed by Shapiro–Wilk test to determine the Normality and Lognormality Tests across groups, and then analyzed by one-way ANOVA with post-hoc Dunnett's multiple comparisons test for groups that had passed the normality test (ns, no significant difference; **P < 0.01). The exact n, SD values, and P values are shown in the data

We also measured Spm content in the transgenic soybean lines. Consistent with Spd, there was no difference in Spm content among the transgenic soybean lines, WT, and CK in the absence of P. sojae challenge (Fig. 3b). The Spm content in transgenic soybean lines was increased to 1.7-fold, 3.4-fold, and 4.0-fold compared with that of WT and CK at 16, 32, and 48 hpi, respectively (Fig. 3b). Upon P. sojae inoculation, all soybean plants overexpressing GmPAO showed a higher accumulation of Spd and Spm as compared to WT and CK.

Spd and Spm inhibit colony growth, zoospore germination, and pathogenicity of P. sojae at high concentrations

Spd and Spm are reported to significantly inhibit conidial proliferation and colony growth of Verticillium dahliae (Mo et al. 2015). To investigate the roles of Spd and Spm during P. sojae infection, we evaluated whether Spd and Spm have inhibitory effects on P. sojae. The application of Spd at 0.7 and 0.8 mg/mL significantly inhibited colony growth in vitro, with application at 0.9 mg/mL fully inhibiting colony growth of P. sojae (Fig. 4a, b). And the application of Spm at 0.2 and 0.3 mg/mL significantly inhibited colony growth in vitro, with application at 0.4 mg/mL extremely inhibiting colony growth (Fig. 4a, b). Of the two PAs tested, Spm displayed a higher P. sojae inhibition activity. Further experiments showed that Spd and Spm could efficiently inhibit zoospore germination (Fig. 4c, d). Spd and Spm also significantly inhibited P. sojae pathogenicity (Additional file 1: Figure S2). Together, these results indicated that high concentrations of Spd and Spm could directly inhibit P. sojae, but the concentrations of Spd and Spm in transgenic lines were far below the inhibitory concentrations. Thus, the increased Spd and Spm in the GmPAO overexpression soybean lines may ensure H₂O₂ accumulation during infection.

Fig. 4

Effects of PAs (Spd and Spm) on colony growth and zoospore germination of P. sojae. a Inhibition of P. sojae hyphal growth by Spd or Spm. Mock represents treatment with ddH₂O. Scale bar, 1 cm. c Inhibition of P. sojae zoospore germination by Spd (1 mg/mL) or Spm (1 mg/mL). Mock represents treatment with ddH₂O. Scale bar, 50 μm. b, d Statistics of colony diameters and zoospore germination rates of P. sojae. Error bar represents mean ± SD, n = 8, n represents sample number. All the experiments were performed three times (biological replicates) with similar results. The data were analyzed by Shapiro–Wilk test to determine the Normality and Lognormality Tests across groups, and then analyzed by one-way ANOVA with post-hoc Dunnett's multiple comparisons test for groups that had passed the normality test (**P < 0.01). The exact n, SD values, and P values are shown in the data

GmPAO significantly enhances resistance against P. sojae in the transgenic soybean lines

To test whether the GmPAO overexpression soybean lines are resistant to P. sojae, we challenged these soybean lines with two representative virulent P. sojae isolates (P7076 and PS14) that belong to distinct genetic clades (Zhang et al. 2019b). First, etiolated hypocotyls were inoculated with equal amounts of mycelia. The results showed that compared with WT and CK, all the GmPAO overexpression lines exhibited significantly reduced lesion length (Fig. 5a, d and Additional file 1: Figure S3) and P. sojae biomass accumulations in hypocotyls (Fig. 5f) after inoculation with either P7076 or PS14. The results preliminarily indicated that all these transgenic soybean lines exhibited substantially enhanced resistance against two P. sojae isolates when compared with WT and CK.

Fig. 5

GmPAO significantly enhances resistance against P. sojae in the transgenic soybean lines. The soybean lines were the same as in Fig. 1. Two P. sojae isolates (P7076 and PS14) that are virulent on WT were used. a, b Representative disease phenotypes at 36 hpi (a) and 60 hpi (b). Hypocotyls or unifoliate leaves were inoculated with P. sojae P7076. Scale bars, 1 cm. c Representative disease phenotypes of seedlings at 12 dpi. The seedlings were inoculated with P. sojae P7076 (5 × 10⁴ zoospores/mL) using the root-dipping inoculation method. Scale bar, 5 cm. d, e Comparisons of the hypocotyl lesion lengths and leaf lesion diameters caused by P. sojae P7076 or P. sojae PS14. Error bar represents mean ± SD, n ≥ 12, n represents sample number. f, g Quantitation of resistance levels via assessment of P. sojae biomass accumulations in inoculated hypocotyls and leaves by genomic qPCR. Error bar represents mean ± SD, n = 3, n represents sample number. h Seedling disease ratings in the transgenic lines over time. The disease symptoms were recorded at 4, 8, and 12 dpi. Disease ratings are shown with different colors (n ≥ 12, n represents sample number). All the experiments were performed three times (biological replicates) with similar results. The data in d–h were analyzed by Shapiro–Wilk test to determine the Normality and Lognormality Tests across groups, and then analyzed by one-way ANOVA with post-hoc Dunnett's multiple comparisons test for groups that had passed the normality test (ns, no significant difference; **P < 0.01). The exact n, SD values, and P values are shown in the data

Second, the soybean leaves were inoculated with equal amounts of mycelia. Compared with WT and CK, all the GmPAO overexpression lines also showed a significant reduction in both lesion size (Fig. 5b, e and Additional file 1: Figure S4) and P. sojae biomass accumulation in leaves (Fig. 5g) after inoculation with either P7076 or PS14. These results also indicated that all these transgenic soybean lines exhibited substantially enhanced resistance compared with WT and CK.

Third, since P. sojae infects roots in the field, we performed a root-dipping inoculation experiment to further evaluate the resistance. Consistent with the results of etiolated hypocotyl and leaf inoculation, all the GmPAO overexpression lines showed enhanced resistance against P. sojae when compared with WT and CK (Fig. 5c, h and Additional file 1: Figure S5). At 12 days post-inoculation (dpi), more than half of the GmPAO overexpression plants exhibited healthy growth, whereas about 80% of WT and CK plants were killed by the pathogen (Fig. 5c, h and Additional file 1: Figure S5). Collectively, these results further indicated that the GmPAO overexpression soybean lines exhibited substantially enhanced resistance compared with WT and CK.

GmPAO enhances resistance to P. capsici and P. parasitica in N. benthamiana

To investigate whether the resistance conferred by GmPAO is broad-spectrum and effective to other Phytophthora pathogens, we infiltrated Agrobacterium tumefaciens cells carrying a GFP or GmPAO fusion construct and then inoculated the infiltrated leaves with P. capsici. The lesion diameters were measured at 36 and 48 hpi. Interestingly, compared with the GFP expression control, the leaves expressing GmPAO exhibited a significant reduction in P. capsici colonization (Fig. 6a, b). Consistent with reduced lesion size, P. capsici biomass was also less in GmPAO-expressing leaves than in the GFP-expressing control (Fig. 6c). Western blot analysis indicated that all the recombinant proteins were properly expressed at expected sizes in N. benthamiana (Fig. 6d). These results indicated that GmPAO enhanced N. benthamiana resistance against P. capsici when overexpressed in planta.

Fig. 6

GmPAO enhances resistance to P. capsici and P. parasitica infections in N. benthamiana. a, e Lesions caused by P. capsici (a) or P. parasitica (e) on N. benthamiana leaves expressing GFP or GmPAO genes. GFP was used as a negative control. Photos were taken at 36 hpi. Scale bars, 1 cm. b, f Comparisons of the lesion diameter caused by P. capsici (b) or P. parasitica (f). Error bar represents mean ± SD, n ≥ 12, n represents sample number. c, g Relative pathogen biomass quantified by genomic qPCR. Error bar represents mean ± SD, n = 3, n represents sample number. d, h Representative immunoblot showing the levels of GFP and GmPAO after transient expression in N. benthamiana leaves. Proteins were stained with Ponceau S to confirm equal loading. All the experiments were performed three times (biological replicates) with similar results. The data in b, c, f, and g were analyzed by Shapiro–Wilk test to determine the Normality and Lognormality Tests across groups, and then analyzed by unpaired t test for groups that had passed the normality test (**P < 0.01). The exact n, SD values, and P values are shown in the data

Furthermore, GFP and GmPAO were transiently expressed in N. benthamiana leaves individually and were subsequently challenged with P. parasitica. The results indicated that GmPAO also significantly improved N. benthamiana resistance against P. parasitica (Fig. 6e–h) in agroinfiltration assays, demonstrating its effectiveness in conferring broad-spectrum Phytophthora resistance. Collectively, these results indicated that GmPAO-mediated polyamine catabolism significantly enhanced plant broad-spectrum resistance to Phytophthora pathogens.

Transgenic soybean lines exhibit no changes in agronomic traits

In crops, genetic immunity to disease often associates with an unintended reduction in growth and yield (Ning et al. 2017). Therefore, we performed detailed agronomic trait tests on these transgenic plants. We first examined the effect of overexpression of GmPAO on morphology, and found that the transgenic lines exhibited no apparent changes in morphology, including growth phenotype, pod, and seed compared with WT and CK (Fig. 7a). We next examined the effect on seed weight and germination rate, and found that there were no significant differences in 100-seed weight and germination rate among the transgenic lines, WT, and CK (Fig. 7b, c). Taken together, we demonstrated that GmPAO confers resistance to multiple Phytophthora pathogens, and no adverse impact on agronomic traits.

Fig. 7

Agronomic traits of GmPAO overexpression soybean lines. The soybean lines were the same as in Fig. 1. a Growth phenotypes (upper) of soybean plants grown in growth chambers for two months. Scale bar, 20 cm. Pod (middle) and Seed (lower) morphology after harvest. Scale bars, 1 cm. b Statistical analysis of 100-seed weight of the transgenic soybean plants. The seeds were randomly chosen and weighed. Error bar represents mean ± SD, n = 20, n represents sample number. c Germination rate of the transgenic soybean plants. The number of germinated seeds was recorded after gemination for 8 days. Error bar represents mean ± SD, n = 3 experiments with 40 seeds. All the experiments were performed three times (biological replicates) with similar results. The data in b and c were analyzed by Shapiro–Wilk test to determine the Normality and Lognormality Tests across groups, and then analyzed by one-way ANOVA with post-hoc Dunnett's multiple comparisons test for groups that had passed the normality test (ns, no significant difference). The exact n, SD values, and P values are shown in the data

Discussion

Phytophthora pathogens cause destructive diseases in agricultural settings, thereby threatening global food security (Wang et al. 2020). Currently, selection and breeding of plant cultivars with broad-spectrum resistance are the most effective approaches for Phytophthora disease management. Compared with conventional breeding, genetic engineering can shorten breeding cycles and facilitate interspecific transfer of resistance genes (Zhou et al. 2021). However, few molecular breeding strategies have been designed to date for Phytophthora disease control or improvement of soybean disease resistance in general. Recent studies showed that H₂O₂ can be generated via Spd and Spm oxidation catalyzed by PAO (Moschou et al. 2009; Mo et al. 2015; Gerlin et al. 2021). Hence, we hypothesized that pathogen-enhanced production of plant Spd and Spm could be oxidized by overexpressed GmPAO to generate H₂O₂ and promote immunity responses, which are specific to pathogen infection but are not triggered in the absence of pathogen challenge.

Previous studies reported that during the interaction of wheat with Fusarium graminearum, the amount of Put in wheat plants increased rapidly while Spd slightly increased and Spm content did not change significantly (Ma et al. 2021). Another research reported that Spm and Put exhibited an increase of 2.5-fold and 2.0-fold, respectively, whereas Spd increased only slightly in the interaction of tobacco with Pseudomonas syringae (Moschou et al. 2009). Upon P. sojae infection, the amounts of Spd and Spm but not Put in soybean were increased (Fig. 1a). These PAs have been demonstrated to be catalyzed by PAO to generate H₂O₂ (Moschou et al. 2009; Seifi and Shelp 2019), which plays a central role in plant defense (Qi et al. 2017; Gao et al. 2021). Based on these results, we proposed that enhancement of Spd and Spm by pathogen infection could be used as targets to ensure pathogen-specific H₂O₂ generation by overexpressing PAO.

Here, we have shown that GmPAO could significantly enhance H₂O₂ content in transgenic soybean lines (Fig. 1d). Moschou et al. (2009) found that overexpression of maize PAO in Nicotiana increased H₂O₂ level. In another example, overexpression of cotton PAO in Arabidopsis also increased H₂O₂ content compared to wild type (Mo et al. 2015). Interestingly, we also found that GmPAO-mediated H₂O₂ only occurred at the P. sojae infection sites (Fig. 1c), thus minimizing cellular damage caused by H₂O₂. Modulation of polyamine metabolism under biotic stress has attracted wide interest; however, the precise functions of these amines in plant defense response against pathogen attack are still poorly understood. Recent studies have demonstrated that plant polyamines are involved in PTI against pathogens (Gerlin et al. 2021). Our results showed that GmPAO-mediated polyamine catabolism promoted flg22/chitin-induced H₂O₂ accumulation (Fig. 2), indicating that GmPAO can also promote PTI. In this disease resistance strategy, GmPAO-mediated H₂O₂ accumulation can improve plant resistance, and GmPAO-promoted PTI can further improve plant resistance (Fig. 8).

Fig. 8

A proposed model illustrating the mechanism of GmPAO-mediated resistance. The two substrates Spd and Spm are increased upon P. sojae infection. GmPAO-mediated Spd/Spm catabolism enhances H₂O₂ production and PTI activation, thereby controlling the infection caused by P. sojae. The activation of PTI response might promote the increase of Spd and Spm contents in soybean

Our results indicated that all the transgenic soybean plants showed higher accumulations of Spd and Spm compared with WT and CK (Fig. 3). This is likely to be a feedback regulation that promotes the accumulation of Spd and Spm in soybean and/or a response of PTI activation (Fig. 8). An interesting question, what is the effect of increased Spd and Spm in soybean on the control of P. sojae? Our results indicated that high concentrations of Spd and Spm could inhibit P. sojae, but the concentrations of Spd and Spm in the transgenic lines were far below the inhibitory concentrations. Thus, the increased Spd and Spm in GmPAO overexpression soybean lines may boost H₂O₂ generation during interaction with pathogens, rather than direct inhibition of P. sojae (Fig. 8).

It has long been established that plants respond to pathogen attack with a transient burst of ROS and that ROS play a central role in plant immune responses (Qi et al. 2017). Meanwhile, ROS are widely produced in different cellular compartments under both biotic and abiotic stress conditions (Qi et al. 2017). Therefore, disease resistance induced by PAO-mediated H₂O₂ production tends to be durable and broad-spectrum. In this study, GmPAO enhanced resistance against P. sojae isolates P7076 and PS14 in the transgenic soybean lines (Fig. 5). Furthermore, we also showed that GmPAO improved N. benthamiana resistance to P. capsici and P. parasitica in agroinfiltration assays (Fig. 6), demonstrating its effectiveness in conferring broad-spectrum resistance. In addition, we measured the agronomic traits of transgenic soybean lines and found no negative effects (Fig. 7). In many cases, constitutively overexpressed transgenes adversely affected plant size and/or seed production. This study provides an approach to express the transgenes only when and where they are needed—at infection sites.

Conclusions

In summary, we develop an efficient metabolic engineering strategy to promote soybean H₂O₂ production upon Phytophthora infection with no adverse impact on agronomic traits. Overexpression of GmPAO may confer broad-spectrum disease resistance by ensuring both PTI activation and H₂O₂ production specifically when and where Phytophthora pathogens are present.

Methods

Plasmid construction and generation of transgenic soybean lines

The sequence of Glycine max polyamine oxidase gene (Glyma.09G227500.1) was obtained from Phytozome databases. The recombinant constructs were introduced into cotyledonary explants of soybean (Glycine max) cultivar Williams 82 by Agrobacterium-mediated transformation (Wang et al. 2020). Soybean transformants were selected using herbicide with glufosinate-ammonium (50 mg/L) as the active ingredient (Bayer CropScience, Research Triangle Park, NC, USA). Transformants were allowed to self-pollinate in growth chamber. After five generations of herbicide selection, T5 homozygous lines were used for the experiments (Lin et al. 2016). Genomic DNA was extracted from leaves of putative transgenic plants and WT. Total RNA was extracted using an RNA-Simple Total RNA Kit (Tiangen, China) according to the manufacturer’s instructions. Soybean cDNAs were synthesized using the Super-ScriptIII First-Strand Kit (Invitrogen, USA). Genomic PCR and quantitative PCR (qPCR) assays were used to screen and validate transgenic lines. GmPAO, selection gene (Bar), and internal reference (ACT11) were amplified using specific primers (Additional file 2: Table S2). All amplicons had the expected sizes.

Plant materials, microbial strains, and growth conditions

Soybean cultivar Williams 82 was grown and maintained in plant growth chambers at an ambient temperature of 25 °C with a 16-h day/8-h night photoperiod. N. benthamiana plants were grown in soil in a plant growth chamber at 25 °C with 60% relative humidity and a 16 h light/8 h dark photoperiod. Two P. sojae isolates (P7076 and PS14) that are virulent to the soybean cultivar Williams 82 were used in this study. P. sojae isolates, P. capsici isolate LT263, and P. parasitica isolate 025 used in this study were routinely cultured at 25 °C in the dark on 10% (v/v) V8 juice medium.

Phytophthora infection assays

P. sojae isolates (P7076 or PS14) that are virulent to Williams 82 were inoculated on soybean etiolated hypocotyls and leaves with mycelia, and seedlings by root-dipping method as described previously (Zhou et al. 2021). Lesion lengths or diameters were measured with the ImageJ software. Inoculated soybean seedlings were incubated for 12 days under controlled environmental conditions (25 °C, 16-h light/8-h dark photoperiod) before phenotypic scoring. The disease symptoms were recorded at 4, 8, and 12 dpi, and the disease rates were classified as no visible disease symptoms, < 50% of lesion areas, > 50% of lesion areas but still alive, and dead plants. The survival rate of WT, CK and transgenic soybean seedings at 12 dpi were statistically analyzed.

For the evaluation of disease resistance by transient expression, N. benthamiana leaves were agroinfiltrated with the binary vector pBIN4 (expressing GFP) on one-half of the leaf and the recombinant pBIN4 vector containing GmPAO was agroinfiltrated on the other half of the same leaf. The infiltrated leaves were inoculated with P. capsici or P. parasitica mycelia at 36 hpi. The lesion diameter was measured at 36/48 hpi, agroinfiltrated leaf samples were collected at 48 hpi and immediately frozen with liquid nitrogen before being stored for Western blot. For evaluation of disease development, relative quantification of P. capsici or P. parasitica biomass in infected leaves was performed by measuring the ratio of pathogen to host genomic DNA by qPCR as previously described (Zhou et al. 2021).

Agrobacterium-mediated transient gene expression

For transient expression assays, transformed Agrobacterium strains were cultured, washed, and resuspended in infiltration buffer (10 mM MgCl₂, 10 mM MES, and 150 μM acetosyringone) to make an appropriate optical density (OD) of 0.3 at 600 nm. Four-week-old N. benthamiana leaves were infiltrated with a 1:1 mixture of resuspended Agrobacterium containing the respective constructs and RNA silencing suppressor P19 (Li et al. 2019a).

SDS–PAGE and Western blot

Proteins from the sample lysate were fractionated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE). Fractionated proteins were electro-transferred from the gel to an Immobilon-PSQ polyvinylidene difluoride membrane using transfer buffer (20 mM Tris, 150 mM glycine). The membrane was then blocked for 30 min at room temperature using phosphate-buffered saline (PBS; pH 7.4) containing 3% nonfat dry milk with shaking at 50 rpm. After washing with PBST (PBS with 0.1% Tween 20), anti-HA (1:2000, Abmart) antibody was added to PBSTM (PBS with 0.1% Tween 20 and 3% nonfat dry milk) and incubated for 90 min. After three times of washing (5 min each) with PBST, the membrane was then incubated with goat anti-mouse IRDye 800CW antibody (Odyssey) at a ratio of 1:10,000 in PBSTM for 30 min. The membrane was finally washed with PBST and visualized with excitations at 700 and 800 nm (Yang et al. 2021).

The UHPLC-MS analysis of PAs

The samples were analyzed by LC–MS system (G2-XS QTof, Waters). 2 μL solution was injected into the UPLC column (2.1 × 100 mm ACQUITY UPLC BEH C18 column containing 1.7 μm particles) with a flow rate of 0.4 mL/min. Buffer A consisted of 0.1% formic acid in water, and buffer B consisted of 0.1% formic acid in acetonitrile. The gradient was 5% Buffer B for 0.5 min, 5–95% Buffer B for 11 min, and 95% Buffer B for 2 min. Mass spectrometry was performed using electrospray source in positive ion mode with MS acquisition mode, with a selected mass range of 100–1200 m/z. The lock mass option was enabled using leucine-enkephalin (m/z 556.2771) for recalibration. The ionization parameters were as follows: 2.5 kV of capillary voltage, 40 V of sample cone, 120 °C of source temperature, and 800 °C of desolvation gas temperature. Data acquisition and processing were performed using Masslynx 4.1 (Mo et al. 2015).

PAO enzymatic activity and H₂O₂ content assays

The level of PAO activity in plant tissues was determined using the spectrophotometric method (Mo et al. 2015). PAO activity was analyzed with a polyamine oxidase assay kit (Cominbio, China) according to the manufacturer’s instructions. Soybean leaves were stained with 1 mg/mL DAB solution for 8 h in the dark at 16 hpi and then decolored with ethanol for light microscopy examination. Samples were equilibrated with 70% (v/v) glycerol for photography using natural light. H₂O₂ content was analyzed with a H₂O₂ assay kit (Beyotime Biotechnology, China) according to the manufacturer’s instructions.

ROS production assay

Microbial pattern-triggered ROS burst assays were performed according to a previous report (Wang et al. 2020) with modifications. In brief, leaf disks were collected from 10-day-old soil-grown soybean plants or 4- to 5-week-old N. benthamiana plants and were incubated in water for 12 h in a 96-well plate. For ROS measurement in N. benthamiana plants, the indicated constructs were transiently expressed in leaves by Agrobacterium-mediated transient expression for 2 days before cutting the leaf disks. The samples were treated with detection buffer, containing 20 μg/mL horseradish peroxidase (Sigma), 20 μM L-012 (Waco), and 1 μM flg22 (Sangon) or 200 μg/mL chitin (Sigma). Light emission was measured at 3-min intervals.

Statistical analysis

Statistical analysis was performed using GraphPad Prism 8.3.0. Shapiro–Wilk test was used to determine normality and lognormality across groups. Groups passing the normality test were compared by unpaired t test or one-way ANOVA with post-hoc Dunnett's test for multiple comparisons (ns, no significant difference; *P < 0.05; **P < 0.01). Results are means ± SD.